topic Re: Unable to assign Pour Points Shapefile in ArcGIS Spatial Analyst Questions

Unable to assign Pour Points Shapefile

SamanArmal1 — Thu, 09 Apr 2015 14:24:02 GMT

Hello everyone,

I need to delineate the Chesapeake Bay watershed. For this reason, I took 16 tiles of DEM production(Arcgrid) from USGS database and merge them through Mosaic Database. I ran spatial analyst toolbox (Fill, Flow Direction, Flow Accumulation respectively) with the same spatial references as that of dem datasets (GCS_North_American_1983). Now, when I want to define pour point Shapefile to finalize watershed delineation, It can not snap the cell with the high accumulation. I am sure that Geographic Coordination System is same for my all layers. But when I want to assign relevant cells into the Shapefile of pour-points, it show a warning in which alert for inconsistency between Shapefile and one of the layers (Attachment-1). Also the result of flow accumulation doesn't seem to be satisfactory (Please see Attachment-2,3).

Is there any help you can give me. I don't know where the problem comes from.

Thank you very much

Re: Unable to assign Pour Points Shapefile

ChrisDonohue__GISP — Thu, 09 Apr 2015 14:37:33 GMT

Some things to check right from the start:

1. What is the Data Frame coordinate system? I ask as the Data frame inherits the coordinate system of the first layer you loaded into the mxd, so may have inherited a different coordinate system than what you want.

2. Are you specifying the coordinate system in the Environments settings when running the tools? Doing so will ensure that the coordinate system remains constant.

Chris Donohue, GISP

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Thu, 09 Apr 2015 14:50:38 GMT

Thank you for reply Chris.

All of my initial ArcGrid DEM data comes from USGS map viewer (USGS TNM 2.0 Viewer ) . They are all in the same format of 1 arc-second <date> 1 x 1 degree ArcGrid, with same Spatial Refrence (GCS_North_American_1983) and the same datum and angular unit. And regarding your second question, yes, I changed cell size (Environment>Raster Analysis) in all the steps, to the "same as" input layer.

Re: Unable to assign Pour Points Shapefile

NeilAyres — Thu, 09 Apr 2015 15:11:59 GMT

How are you getting your pour points and how do you define the snapping distance given that this data is in degrees.

Personally, if I was doing this, I would put the data into a flat projected system first for the analysis.

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Thu, 09 Apr 2015 15:22:19 GMT

Unfortunately, It doesn't allow me to get/snap pour points. The warning in the first image pops out when I first add the new shapefile for pour points. By the way, is there any link in which I can learn about flat projected system that you mentioned?

Re: Unable to assign Pour Points Shapefile

NeilAyres — Thu, 09 Apr 2015 15:34:38 GMT

Map projections beginning :

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

What do you mean "add pour points". I was asking how are you obtaining your pour points for the purposes of a watershed deliniation. That is all in the help files under hydrology.

Start here :

ArcGIS Help (10.2, 10.2.1, and 10.2.2)

If you are simply adding a point shapefile which you have created in ArcCatalog. Then there is definitely something different about its coordinate system relative to the DEM coord sys. Right click on it and find out what it is defined as.

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Thu, 09 Apr 2015 15:43:06 GMT

Oh, I am trying to add the point manually (through visual inspection, by clicking in the center of the high flow accumulation cells to objectify the outlets). Sorry for misunderstanding. I am sure that all my layers are in the same coordination system. When I add new shape file in ArcCatalog, I just simply change the co-system to what I have for other layers. Is there any other thing should I do?

Re: Unable to assign Pour Points Shapefile

ChrisDonohue__GISP — Thu, 09 Apr 2015 15:55:25 GMT

As Neil pointed out, your data currently is in a Geographic Coordinate System (GCS). However, for what you want to do, a Projected Coordinate System is the way to go.

Here's how I recommend you proceed:

The first step then is to understand projected coordinate systems. Start with the links that Neil provided.

Then choose an appropriate projected coordinate system - one that works well for spatial extent of your project and minimizes distortion for your analysis.

Then there is the data preparation stage, where all your data is reprojected to the chosen system.

Once all this is done, you can proceed with your analysis. This all may sound like a lot of work, but it is well worth it to get solid results and to avoid issues like you have been encountering.

Chris Donohue, GISP

Re: Unable to assign Pour Points Shapefile

NeilAyres — Thu, 09 Apr 2015 18:31:14 GMT

Well, once you have decided upon the correct projected coordinate system for your data area, you should really start again.

Just project the actual DEM, then start again generating the filled dem, flow direction etc. Projecting from GCS (or from anything to anything) can actually "close" flow lines where they were open before depending on the DEM resolution and the resampling (in the projection step) involved.

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Thu, 09 Apr 2015 23:36:15 GMT

I found the proper Projection system in the table provided by ESRI help (http://resources.arcgis.com/en/help/main/10.2/003r/pdf/projected_coordinate_systems.pdf), which in this case is (WKID=32618, WGS_1984_UTM_Zone_18N). I merged DEM tiles in mosaic dataset with this projectcion system and ran Fill, Flow-Direction and Flow-Accumulation tools in the same projection system, Each time I set the cell size under Environment>Raster Analyze as same as that of input layer. Unfortunately, I still have the same problem and I am not able to snap the pour points. I think the problem comes from a bay in my area of study and the challenge is how to rout the flow in there.

Re: Unable to assign Pour Points Shapefile

SepheFox — Fri, 10 Apr 2015 01:27:01 GMT

Can you post a screenshot of your flow direction? Your flow accumulation does look weird.

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Fri, 10 Apr 2015 01:48:55 GMT

Yes, I think its weird either, here is a Flow-direction result:

Re: Unable to assign Pour Points Shapefile

SepheFox — Fri, 10 Apr 2015 01:55:44 GMT

Yes, something is definitely wrong with it. Do you mind posting a screenshot of the DEM?

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Fri, 10 Apr 2015 02:07:49 GMT

Not at all. Here is a original DEM datasets, taken from (USGS TNM 2.0 Viewer), with Geographic Coordination of GCS_North_American_1983 and Cell size of 1-Arc second:

This is the Mosaic Dataset, as s result of the DEM tiles merging, with projection Coordination system of (WGS_1984_UTM_Zone_18N).

Thank you,

Re: Unable to assign Pour Points Shapefile

SepheFox — Fri, 10 Apr 2015 02:22:22 GMT

Hmmm, well, that looks reasonable, although it's hard to tell down in the flat areas. Something is not coming out right with that flow direction though. Maybe try running it again.

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Fri, 10 Apr 2015 02:31:16 GMT

Thank you so much Mr.Fox. Don't you think the problem comes from a bay in my area of study (Chesapeake Bay watershed)? I have also a bathymetric DEM datasets of watershed. Do you have any idea if they will be helpful to rout the flow to the bay? I ran the flow_direction and Flow_Accumulation at least three times, each time at least 2.5 hours. Still not working.

Re: Unable to assign Pour Points Shapefile

SepheFox — Fri, 10 Apr 2015 02:35:39 GMT

Well, large flat expanses can do weird things, but it looks just as odd at higher elevations. Out of curiosity, when you open the raster properties of your projected DEM, what does it say for cell size?

Re: Unable to assign Pour Points Shapefile

SamanArmal1 — Fri, 10 Apr 2015 02:41:04 GMT

The cell size is same as that of original DEMs: 30.92208078, 30.92208078.